In [1]:
import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib
import matplotlib.pyplot as plt
import pickle as pkl

from scipy.stats import ttest_ind
from sklearn.preprocessing import RobustScaler
from sklearn.manifold import TSNE
from sklearn.mixture import GaussianMixture

font= matplotlib.font_manager.FontProperties(fname= 'C:\Windows\\Fonts\\조선일보명조.ttf').get_name()
matplotlib.rc('font', family=font)

데이터를 로드하고 분석에 적합한 형태로 변환함

In [2]:
df_bymonth= pd.read_csv('./Pre_processed_data/groupby_ID_YEARMONTH.csv')
In [3]:
pt_bytsne= pd.read_pickle('./pkl_result/data_tSNE_dim2_perplex30.pkl')
In [4]:
df_raw= pd.read_csv('./Pre_processed_data/merged_dataframe_shopping.csv', delimiter= ',', encoding= 'euc-kr')
In [5]:
df_grouped= df_raw.sort_values('ID')
In [6]:
agg_func= {'RCT_NO': 'count', 'BIZ_UNIT': lambda x: x.nunique(), 'PD_S_C': lambda x: x.nunique(), \
           'BR_C': lambda x: x.nunique(), 'BUY_AM': 'sum'}
In [7]:
df_grouped= df_grouped.groupby('ID').agg(agg_func).reset_index(drop= False)
In [8]:
df_demo= df_raw[['ID', 'AGE_PRD', 'IS_MALE', 'IS_FEMALE', 'PRVIN']].drop_duplicates(subset= 'ID').sort_values('ID').reset_index(drop= True)
In [9]:
df_byitem= pd.read_pickle('./pkl_result/item_selected_by_middleline.pkl')
In [10]:
df_byitem= df_byitem.reset_index(drop=False)
In [11]:
df_item_consumed= pd.read_pickle('./pkl_result/item_selected_by_middleline.pkl')
In [12]:
df_item_consumed.reset_index(drop= False, inplace= True)
In [13]:
def plot_results(X, Y_= None):
    try:
        if Y_.any():
            for i in range(Y_.nunique()):
                grp= sns.regplot(x= X.loc[Y_==i, 0], y= X.loc[Y_==i, 1], fit_reg= False, scatter_kws= {'s': 2})
        plt.show()
    except AttributeError:
        grp= sns.regplot(x= X.iloc[:,0], y= X.iloc[:, 1], fit_reg= False, scatter_kws= {'s': 2})
        plt.show()
In [14]:
def plot_clusters(X, Y_):
    for i in sorted(Y_.unique().tolist()):
        grp= sns.regplot(x= X.iloc[:,0], y= X.iloc[:, 1], fit_reg= False, scatter_kws= {'s': 2})
        grp= sns.regplot(x= X.loc[Y_==i, 0], y= X.loc[Y_==i, 1], fit_reg= False, scatter_kws= {"color": "red", 's': 2})
        plt.show()
In [15]:
prefered_item_dict_05= {}
ft_dict= {}
def find_prefered_item(df, label, p_val= 0.05):
    for i in sorted(label.unique().tolist()):
        ft_dict[i]= df.loc[label==i, :]
        print('\nfor cluster {}:\n'.format(str(i)))
        prefered_item_dict_05[i]= []
        for item in df.columns.values.tolist():
            (stat_item, p_val_item)=  ttest_ind(ft_dict[i][item], df[item], equal_var= False)
            if p_val_item< p_val and stat_item>0:
                print('item {} has different mean with total dataset; ({}, {})'.format(item, stat_item, p_val_item))
                prefered_item_dict_05[i].append(item)
                
def find_prefered_by_clusters(df, label, i, j, p_val= 0.05):
    for item in df.columns.values.tolist():
        (stat_item, p_val_item)=  ttest_ind(df.loc[label==i, item], df.loc[label==j, item], equal_var= False)
        if p_val_item< p_val:
            print('item {} has different mean with total dataset; ({}, {})'.format(item, stat_item, p_val_item))
In [16]:
plot_results(pt_bytsne[[0,1]], pt_bytsne['y_14_clst'])
In [17]:
plot_results(pt_bytsne[[0,1]])
In [18]:
plot_clusters(pt_bytsne[[0,1]], pt_bytsne['y_14_clst'])
In [19]:
plot_clusters(pt_bytsne[[0,1]], df_demo['IS_FEMALE'])
#성별이 클러스터와 연관이 있는 인구통계 정보인지 확인함. 아래의 시각화 결과에 따르면, 성별은 클러스터를 구분짓는 요소는 아니지만 클러스터 내에서 소비 품목의 차이는 드러났음.
In [20]:
plot_clusters(pt_bytsne[[0,1]], df_demo['AGE_PRD'])
# 연령대의 경우도 마찬가지로 연령대와 관계 없이 폭넓게 분포하고 있었음.

성별, 연령대 또는 나이에 따라 사용자를 나누어 클러스터링해 보았음. 아래는 여성 사용자만을 대상으로 클러스터링한 결과임

In [21]:
pt_bytsne= pd.concat([df_demo['ID'], pt_bytsne], axis= 1)
In [22]:
pt_female= pt_bytsne.loc[df_demo['IS_FEMALE']==1, :].reset_index(drop= True)
In [23]:
pt_female.head()
Out[23]:
ID 0 1 y_14_clst y_28_clst y_24_clst
0 2 53.016911 -32.975819 10 8 8
1 3 -22.183092 52.640892 3 9 9
2 6 42.187847 32.573856 5 17 17
3 7 -27.630230 52.580269 3 12 12
4 8 -5.850988 27.713455 12 3 3
In [24]:
pt_male= pt_bytsne.loc[df_demo['IS_FEMALE']==0, :].reset_index(drop= True)
In [25]:
label_dict= {}
for i in range(10, 41, 2):
    gmm = GaussianMixture(n_components=i, covariance_type='full', random_state= 42).fit(pt_female[[0,1]])
    label_dict[(2,i)]= pd.Series(gmm.predict(pt_female[[0,1]]))
    plot_results(pt_female[[0,1]], label_dict[(2,i)])
In [26]:
gmm = GaussianMixture(n_components=14, covariance_type='full', random_state= 42).fit(pt_female[[0,1]])
label_dict[(2,i)]= pd.Series(gmm.predict(pt_female[[0,1]]))
plot_results(pt_female[[0,1]], label_dict[(2,14)])

# 여성들만을 군집화한 경우에도 유사한 클러스터가 형성되는 것을 확인함
In [27]:
plot_clusters(pt_female[[0,1]], label_dict[(2,14)])
In [28]:
df_female= df_item_consumed.loc[df_demo['IS_FEMALE']==1, :].drop('ID', axis= 1).reset_index(drop= True)
In [29]:
df_female.head()
Out[29]:
H&B선물세트 VIDEOGAME 가공식품 가공우유 가구 가방브랜드 계절완구 고급 고양이용품 골프 ... 필기용품 한방차 한우선물세트 해초 헬스용품 호주산소고기 홍인삼 화과자 황태 훼이셜케어
0 0.0 0.0 6.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 ... 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
1 0.0 0.0 3.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ... 0.0 0.0 0.0 0.0 0.0 4.0 0.0 0.0 0.0 0.0
2 0.0 0.0 2.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ... 0.0 0.0 0.0 0.0 0.0 14.0 0.0 1.0 0.0 0.0
3 0.0 0.0 18.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ... 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
4 0.0 0.0 23.0 2.0 0.0 0.0 0.0 0.0 0.0 0.0 ... 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

5 rows × 149 columns

In [30]:
find_prefered_item(df_female, label_dict[(2,14)])

for cluster 0:


for cluster 1:

item 고양이용품 has different mean with total dataset; (3.299518677371703, 0.0009909005140444082)
item 골프 has different mean with total dataset; (2.2027005968132984, 0.027740297838069142)
item 기능성건강 has different mean with total dataset; (3.4992108669368958, 0.0004793779071884237)
item 남성어덜트캐주얼 has different mean with total dataset; (2.0028544157994808, 0.04534364652717147)
item 석류 has different mean with total dataset; (2.090670045665515, 0.03670272271987567)
item 스낵 has different mean with total dataset; (16.74676631306228, 1.558399904946653e-58)
item 스포츠슈즈 has different mean with total dataset; (1.9651539914316039, 0.04955514830267115)
item 시즌행사 has different mean with total dataset; (3.277927144713126, 0.0010671906320496205)
item 시티캐주얼 has different mean with total dataset; (3.26787436537591, 0.0011070949867525788)
item 언더웨어행사 has different mean with total dataset; (2.1998402667187427, 0.027943694964898326)
item 여성양말 has different mean with total dataset; (2.1488676084319867, 0.03177566201141486)
item 여성캐주얼브랜드 has different mean with total dataset; (2.2998924911678675, 0.021554416225706415)
item 유가공 has different mean with total dataset; (4.489761869189406, 7.639010897485679e-06)
item 유아용품 has different mean with total dataset; (4.969825562592848, 7.440818348791663e-07)
item 육포 has different mean with total dataset; (2.61715158492918, 0.008942899383670325)
item 이지캐주얼브랜드 has different mean with total dataset; (9.116678446307832, 1.698785084427786e-19)
item 일반욕실용품 has different mean with total dataset; (3.1315753697371673, 0.0017668736069170812)
item 일반헤어케어 has different mean with total dataset; (5.320540531255847, 1.1757586014716487e-07)
item 자동차방향제/소취제 has different mean with total dataset; (2.3633501446683924, 0.018218991815612654)
item 자몽 has different mean with total dataset; (3.4961957674164825, 0.000483965223227956)
item 전통주 has different mean with total dataset; (2.7262226191205965, 0.0064353464576787185)
item 조기/명태 has different mean with total dataset; (2.7844816550654947, 0.005421942723645665)
item 조리반찬 has different mean with total dataset; (5.829384141093576, 6.69317988462707e-09)
item 즉석구이김 has different mean with total dataset; (3.3292939653298985, 0.0008908003285027032)
item 즉석구이안주 has different mean with total dataset; (2.761922559053466, 0.005812776187761255)
item 즉석도정미 has different mean with total dataset; (2.0201197872278196, 0.04352007894963683)
item 차 has different mean with total dataset; (3.2819222791706495, 0.0010505744017778357)
item 천연조미료 has different mean with total dataset; (2.1598378912790523, 0.030918314136866885)
item 캐주얼슈즈 has different mean with total dataset; (2.708985962323969, 0.006819355442602371)
item 튀김 has different mean with total dataset; (8.738288771106687, 5.397328154161333e-18)
item 특산물행사 has different mean with total dataset; (2.0210336406477367, 0.04343083626531548)
item 파인애플 has different mean with total dataset; (3.2077549690916483, 0.001357670990579247)
item 풋/핸드케어 has different mean with total dataset; (1.9937408385161055, 0.046334013199781296)
item 피자 has different mean with total dataset; (2.7535758197210662, 0.005961707314918512)
item 해초 has different mean with total dataset; (2.7304216619233164, 0.006391809829774491)
item 호주산소고기 has different mean with total dataset; (6.486466114885406, 1.0794459720056244e-10)
item 홍인삼 has different mean with total dataset; (2.4126080885119414, 0.01594407706362072)
item 황태 has different mean with total dataset; (4.036475031711779, 5.67086854421223e-05)
item 훼이셜케어 has different mean with total dataset; (4.088110433896342, 4.530666881109829e-05)

for cluster 2:

item 가공식품 has different mean with total dataset; (4.169979561618935, 4.004112980683585e-05)
item 골프용품 has different mean with total dataset; (2.4242858615716805, 0.01589207665766375)
item 모피/피혁 has different mean with total dataset; (2.227991389994793, 0.026646212747087814)
item 시티웨어 has different mean with total dataset; (11.189041721371742, 2.2857928495627204e-24)

for cluster 3:


for cluster 4:

item 가공식품 has different mean with total dataset; (4.296338237349972, 1.9633809882508828e-05)
item 가공우유 has different mean with total dataset; (22.67109618660454, 1.65571374758071e-98)
item 고급 has different mean with total dataset; (3.5327085210484697, 0.0004387078204510708)
item 너트류 has different mean with total dataset; (3.808606598123489, 0.00015253606009145372)
item 봉투보증금 has different mean with total dataset; (3.2267445173992773, 0.0013144214563215967)
item 샐러드 has different mean with total dataset; (2.015783962142943, 0.04421796517624327)
item 스타킹 has different mean with total dataset; (2.349965170780579, 0.019061933798275164)
item 파인애플 has different mean with total dataset; (3.146907089735294, 0.0017252174898439315)
item 화과자 has different mean with total dataset; (2.0913603854854945, 0.03679823077036847)

for cluster 5:

item VIDEOGAME has different mean with total dataset; (3.1391533791016957, 0.0017561419401339127)
item 가방브랜드 has different mean with total dataset; (2.0692131063237613, 0.03883144370506655)
item 구기용품 has different mean with total dataset; (3.229625148388257, 0.0012881411350066146)
item 드레스화브랜드 has different mean with total dataset; (2.405731873237307, 0.016219144139619626)
item 라켓스포츠 has different mean with total dataset; (2.900300427611405, 0.0038272432181782534)
item 레스토랑 has different mean with total dataset; (5.722123845027699, 1.4689457413635092e-08)
item 문화센터 has different mean with total dataset; (3.7450904876263262, 0.0001933301482161954)
item 미시캐주얼 has different mean with total dataset; (2.6622513506477214, 0.007908946029445441)
item 복합샵 has different mean with total dataset; (5.502631110923063, 5.0599555411693214e-08)
item 브랜드침구 has different mean with total dataset; (3.3261214885453922, 0.000920236262653159)
item 비즈니스정장 has different mean with total dataset; (2.795658478086265, 0.005301426313026446)
item 소가구 has different mean with total dataset; (2.286190630863014, 0.022497269183683426)
item 수예시즌 has different mean with total dataset; (2.572266051425233, 0.01027556229791486)
item 수족관용품 has different mean with total dataset; (3.059956551400694, 0.002282546007111554)
item 스낵 has different mean with total dataset; (3.0507958267608606, 0.0023472333835642145)
item 스포츠ACC has different mean with total dataset; (2.3808897998518184, 0.017497488868674586)
item 스포츠슈즈 has different mean with total dataset; (3.7638729570974294, 0.00017904877543202676)
item 시즌슈즈 has different mean with total dataset; (2.3449346538540445, 0.019266490583197918)
item 시즌잡화 has different mean with total dataset; (3.0572784112854094, 0.0023062866662915157)
item 시즌행사 has different mean with total dataset; (5.190552408178539, 2.638680518768235e-07)
item 시티캐주얼 has different mean with total dataset; (2.684460687395138, 0.0073984492121600244)
item 아동브랜드 has different mean with total dataset; (12.051141219593685, 7.895592509276176e-31)
item 아동침구 has different mean with total dataset; (3.344648842338019, 0.000861004775779973)
item 양말선물세트 has different mean with total dataset; (2.2127278518412683, 0.027179905070099757)
item 언더웨어행사 has different mean with total dataset; (3.5382266267707227, 0.00042581547153327945)
item 여성양말 has different mean with total dataset; (4.586106799824049, 5.2264246351596904e-06)
item 여성캐주얼브랜드 has different mean with total dataset; (5.41677512014305, 8.057757139423002e-08)
item 우산/양산 has different mean with total dataset; (4.0755348782446195, 5.039266538266879e-05)
item 원예용품 has different mean with total dataset; (2.771136471693107, 0.005714618754988028)
item 윈터스포츠 has different mean with total dataset; (1.976133588091999, 0.048480968672344145)
item 유아식기 has different mean with total dataset; (5.987745863360044, 2.9591378635052835e-09)
item 유아토들러 has different mean with total dataset; (4.4030318504429005, 1.2097409404265626e-05)
item 음반영상 has different mean with total dataset; (2.477014149070591, 0.01345286352357941)
item 이온음료 has different mean with total dataset; (4.1772835929693075, 3.262227283015938e-05)
item 이유식/유아간식 has different mean with total dataset; (2.9400017137169403, 0.003367972943860945)
item 이지캐주얼브랜드 has different mean with total dataset; (3.516266350869233, 0.0004589154658185845)
item 일반욕실용품 has different mean with total dataset; (6.570630521599731, 8.823130833759374e-11)
item 일반헤어케어 has different mean with total dataset; (2.4370896988412127, 0.015007177285403264)
item 자동차방향제/소취제 has different mean with total dataset; (3.8122501574926604, 0.00014773766941317146)
item 자전거 has different mean with total dataset; (2.253100425479823, 0.024515160430885093)
item 잡화멀티샵 has different mean with total dataset; (3.6862576540256966, 0.00024230594700336156)
item 장갑 has different mean with total dataset; (4.075172723625158, 5.049863395889176e-05)
item 장식용품 has different mean with total dataset; (2.895350285831641, 0.0038858733981311762)
item 주니어양말 has different mean with total dataset; (3.848673031991668, 0.00012806047411557665)
item 준보석 has different mean with total dataset; (2.2866872216273717, 0.02246780422280091)
item 즉석구이김 has different mean with total dataset; (3.438455499031144, 0.0006135635148857926)
item 즉석구이안주 has different mean with total dataset; (2.7838558310824024, 0.005485070230344183)
item 즉석도정미 has different mean with total dataset; (2.923474038100338, 0.0035515019681050535)
item 차 has different mean with total dataset; (4.075877350494336, 4.9897533684391174e-05)
item 천연조미료 has different mean with total dataset; (2.5787901103295394, 0.010088916515979544)
item 카시트/매트 has different mean with total dataset; (2.362912259396963, 0.018365691423449376)
item 카페 has different mean with total dataset; (2.197175939892018, 0.028260200962921186)
item 캐주얼슈즈 has different mean with total dataset; (3.135721333843031, 0.001775884512828499)
item 캐주얼화브랜드 has different mean with total dataset; (3.1452972715583463, 0.0017109963662398048)
item 커튼류 has different mean with total dataset; (2.8183220942025846, 0.004944179149896703)
item 튀김 has different mean with total dataset; (2.9270362251739055, 0.003512773508559976)
item 판매시설 has different mean with total dataset; (5.193152059614473, 2.6271001266393263e-07)
item 편의시설 has different mean with total dataset; (5.4988240839221705, 5.154387564364852e-08)
item 푸드코트 has different mean with total dataset; (13.291866607767933, 1.5532551958740325e-36)
item 헬스용품 has different mean with total dataset; (2.0774775063512476, 0.03805282813787264)
item 홍인삼 has different mean with total dataset; (2.2437765465853268, 0.02510429246746001)
item 황태 has different mean with total dataset; (2.6692733172703442, 0.0077491241466846805)

for cluster 6:

item H&B선물세트 has different mean with total dataset; (2.19468540966767, 0.028696471663472816)
item 가공식품 has different mean with total dataset; (8.896651243111338, 1.255737125736185e-17)
item 가공우유 has different mean with total dataset; (18.462679372883937, 8.064887538563822e-57)
item 계절완구 has different mean with total dataset; (2.305311501744621, 0.021611523046977124)
item 고급 has different mean with total dataset; (4.070322543613183, 5.558716533108937e-05)
item 과일야채 has different mean with total dataset; (2.9194502193580956, 0.003681355848781106)
item 구기용품 has different mean with total dataset; (2.5756948795244514, 0.010324229097019126)
item 기능성건강 has different mean with total dataset; (2.3633458798884615, 0.018526844728952444)
item 너트류 has different mean with total dataset; (2.7553266710329014, 0.006102661157002396)
item 레스토랑 has different mean with total dataset; (2.7622243213812046, 0.005969129292330138)
item 롤샌드/핫도그 has different mean with total dataset; (2.7321596300501723, 0.006544440509741406)
item 리쿼 has different mean with total dataset; (4.016964811084675, 6.927210349913465e-05)
item 매트 has different mean with total dataset; (2.496104268418408, 0.01291385195110576)
item 봉투보증금 has different mean with total dataset; (4.537041744327637, 7.355697128410409e-06)
item 샐러드 has different mean with total dataset; (2.022806413000462, 0.04369345602563288)
item 생선초밥 has different mean with total dataset; (2.6172930712903146, 0.009163474095521016)
item 수입게 has different mean with total dataset; (2.1000942816404233, 0.036281759659578965)
item 수족관용품 has different mean with total dataset; (3.2560542121921814, 0.001214724355353975)
item 스낵 has different mean with total dataset; (7.36318633690931, 8.514853996271239e-13)
item 스타킹 has different mean with total dataset; (5.988177534343326, 4.3396922642130456e-09)
item 스포츠슈즈 has different mean with total dataset; (2.233675043116174, 0.025991011505106677)
item 시즌행사 has different mean with total dataset; (4.4540756311072585, 1.062349812210856e-05)
item 시티캐주얼 has different mean with total dataset; (1.9964644625942678, 0.046452551544514255)
item 식빵 has different mean with total dataset; (3.167068025053777, 0.001646276301261321)
item 아동브랜드 has different mean with total dataset; (2.835693965104684, 0.004773143844169801)
item 아동완구 has different mean with total dataset; (2.7181717064138744, 0.006823158656180564)
item 언더웨어행사 has different mean with total dataset; (2.2652591724076183, 0.023965302909281904)
item 여성양말 has different mean with total dataset; (2.24634072442009, 0.02515286875852458)
item 우산/양산 has different mean with total dataset; (2.412642908237124, 0.016235793013471566)
item 유가공 has different mean with total dataset; (3.9023946745881886, 0.00010955913382563757)
item 육포 has different mean with total dataset; (3.776652514243242, 0.00018059702762960515)
item 음반영상 has different mean with total dataset; (1.9975648609894918, 0.04635577779725012)
item 이온음료 has different mean with total dataset; (4.228187519355287, 2.8615007144877903e-05)
item 이유식/유아간식 has different mean with total dataset; (2.2758965663843975, 0.023301981315638442)
item 일반욕실용품 has different mean with total dataset; (3.6747322891027645, 0.0002663583648462336)
item 일반헤어케어 has different mean with total dataset; (3.6504897772714218, 0.00029165189371464117)
item 자몽 has different mean with total dataset; (2.657227941058873, 0.00815917058525223)
item 장식용품 has different mean with total dataset; (2.002492144068149, 0.0458216160079234)
item 제습제 has different mean with total dataset; (4.35083932450495, 1.6778149519934578e-05)
item 조리반찬 has different mean with total dataset; (2.959884442627607, 0.003241477885378382)
item 준보석/시계 has different mean with total dataset; (2.428077683590707, 0.015565884811163638)
item 즉석구이김 has different mean with total dataset; (2.0377490731020953, 0.0421228559344189)
item 즉석구이안주 has different mean with total dataset; (3.121500186542463, 0.0019142172146312587)
item 즉석도정미 has different mean with total dataset; (3.8957381914056737, 0.00011250734575521332)
item 차 has different mean with total dataset; (4.476494665495747, 9.651964543803694e-06)
item 커피선물세트 has different mean with total dataset; (2.012227402419698, 0.04479850585794824)
item 탈취제 has different mean with total dataset; (2.6850038696987832, 0.007437365248417835)
item 튀김 has different mean with total dataset; (3.5644959603735913, 0.00040374738087551253)
item 특산물행사 has different mean with total dataset; (2.956456309658245, 0.0032755036153706393)
item 파인애플 has different mean with total dataset; (3.5998614780999914, 0.0003530943176306138)
item 편의시설 has different mean with total dataset; (2.629746646171583, 0.00883110128828684)
item 편의점카페 has different mean with total dataset; (2.698784390218705, 0.007225818684432439)
item 포장반찬HMR has different mean with total dataset; (2.0430638150279155, 0.0416271722613942)
item 푸드코트 has different mean with total dataset; (3.2304433114980236, 0.0013208979888890827)
item 풋/핸드케어 has different mean with total dataset; (2.871370048642461, 0.004277262401952289)
item 프리미엄 has different mean with total dataset; (2.5486550022831826, 0.011151919207580317)
item 피자 has different mean with total dataset; (2.4939003520921585, 0.012997006688289733)
item 피혁잡화 has different mean with total dataset; (3.3113970201935667, 0.0009829417032764217)
item 필기용품 has different mean with total dataset; (3.050994937144037, 0.002418001852558694)
item 호주산소고기 has different mean with total dataset; (4.8317814129719485, 1.8455840176034636e-06)
item 홍인삼 has different mean with total dataset; (2.0351129175260128, 0.04242325460910611)
item 화과자 has different mean with total dataset; (3.425935634345291, 0.0006701364732934837)
item 황태 has different mean with total dataset; (2.498320696131178, 0.012819456318517846)
item 훼이셜케어 has different mean with total dataset; (3.066738664971315, 0.0022942636343456896)

for cluster 7:

item 브랜드침구 has different mean with total dataset; (2.041790380942419, 0.041390063001026434)
item 수족관용품 has different mean with total dataset; (4.207645484708913, 2.7799488292674238e-05)
item 유아식기 has different mean with total dataset; (2.621717126641919, 0.008831305976674746)
item 유아토들러 has different mean with total dataset; (3.464117668536925, 0.0005510184878053284)
item 자동차방향제/소취제 has different mean with total dataset; (2.403860586829403, 0.016369548979066912)
item 차 has different mean with total dataset; (2.7460415678363406, 0.0061197335874611245)
item 편의시설 has different mean with total dataset; (5.441985897135, 6.37633507752101e-08)

for cluster 8:

item 주유소 has different mean with total dataset; (4.684355416880827, 4.037087891189402e-06)

for cluster 9:

item 가공식품 has different mean with total dataset; (16.576842168928824, 1.5253742326741789e-56)
item 가구 has different mean with total dataset; (4.205197567169471, 2.7764552566674702e-05)
item 고급 has different mean with total dataset; (2.6183205858883465, 0.008932593098766227)
item 남성정장 has different mean with total dataset; (8.507735905962754, 4.482250170627378e-17)
item 대형가전 has different mean with total dataset; (4.628566871489823, 4.027773188080562e-06)
item 보석 has different mean with total dataset; (2.7703526885035044, 0.005671559639356199)
item 아이메이크업 has different mean with total dataset; (5.2566043400674225, 1.68113860022173e-07)
item 준보석/시계 has different mean with total dataset; (4.31772655770003, 1.686833160510731e-05)
item 피혁잡화 has different mean with total dataset; (5.046317386661228, 5.112021680827198e-07)

for cluster 10:

item H&B선물세트 has different mean with total dataset; (2.336007559538318, 0.01966790557602582)
item 가공식품 has different mean with total dataset; (4.30559515212599, 1.8071835992884385e-05)
item 가공우유 has different mean with total dataset; (3.088726765048904, 0.00204390546823889)
item 가방브랜드 has different mean with total dataset; (3.0209159188603585, 0.002580196927794116)
item 고양이용품 has different mean with total dataset; (2.0780203679345783, 0.03789654297534873)
item 골프 has different mean with total dataset; (2.438083244685053, 0.014925823769131754)
item 구기용품 has different mean with total dataset; (2.5230305212432445, 0.011776037508069222)
item 기능성건강 has different mean with total dataset; (4.851036904876932, 1.4151792993014777e-06)
item 김선물세트 has different mean with total dataset; (2.101713596417747, 0.035806761110779614)
item 동물병원 has different mean with total dataset; (3.357156762527558, 0.0008166152010553196)
item 드레스화브랜드 has different mean with total dataset; (2.453913060962311, 0.014195827194611198)
item 레스토랑 has different mean with total dataset; (5.30933387233083, 1.3353566915805156e-07)
item 문화센터 has different mean with total dataset; (2.2708125908280343, 0.023341688551382432)
item 미시캐주얼 has different mean with total dataset; (2.659983182372438, 0.007937515042047857)
item 벽지류 has different mean with total dataset; (3.01989550540186, 0.002589688034272541)
item 브랜드침구 has different mean with total dataset; (2.974139472785427, 0.003004144545314632)
item 비즈니스정장 has different mean with total dataset; (3.5986385648112393, 0.00033460271940092336)
item 생선초밥 has different mean with total dataset; (2.4996605100954135, 0.01258758357359572)
item 석류 has different mean with total dataset; (3.5590827659700213, 0.0003890466375662874)
item 소가구 has different mean with total dataset; (3.724767655575817, 0.00020584042080109538)
item 수예시즌 has different mean with total dataset; (4.823207731161628, 1.6233255376231674e-06)
item 수입게 has different mean with total dataset; (2.6130632326403855, 0.00909879978932009)
item 수입돼지고기 has different mean with total dataset; (2.5087109565887524, 0.012272491944525558)
item 수족관용품 has different mean with total dataset; (3.7887353230010743, 0.00015976657348003108)
item 스낵 has different mean with total dataset; (14.634684525390055, 3.2819016341297396e-44)
item 스타킹 has different mean with total dataset; (5.363247156762912, 1.0060050635852453e-07)
item 스포츠ACC has different mean with total dataset; (2.3070601563637174, 0.021246955130024742)
item 스포츠슈즈 has different mean with total dataset; (3.1217481305278807, 0.0018458601854104834)
item 스포츠의류 has different mean with total dataset; (2.206296767296592, 0.027569949416031157)
item 시계 has different mean with total dataset; (2.693997425891744, 0.007170131008612731)
item 시즌슈즈 has different mean with total dataset; (2.7068421636788953, 0.006899624161071093)
item 시즌잡화 has different mean with total dataset; (2.6989801386323733, 0.007061670696631192)
item 시즌행사 has different mean with total dataset; (5.028888818883001, 5.79558303879999e-07)
item 시티캐주얼 has different mean with total dataset; (4.279497025767715, 2.0399662114201237e-05)
item 식사류 has different mean with total dataset; (5.715436361270091, 1.4382782653032952e-08)
item 식자재 has different mean with total dataset; (2.9630226548756355, 0.0031171816376912207)
item 아동브랜드 has different mean with total dataset; (2.2368399892840345, 0.0254934487638671)
item 아동침구 has different mean with total dataset; (3.420257749584599, 0.0006497221396002999)
item 아로마 has different mean with total dataset; (4.919178689825357, 1.0095099755165663e-06)
item 아이메이크업 has different mean with total dataset; (2.2294131334354232, 0.02599680663256399)
item 애견용품 has different mean with total dataset; (9.09860044916775, 4.725201467336431e-19)
item 언더웨어행사 has different mean with total dataset; (4.098772854837724, 4.463065875182085e-05)
item 여성양말 has different mean with total dataset; (5.6897530342605975, 1.6437314917758052e-08)
item 여성캐주얼브랜드 has different mean with total dataset; (4.598210697763472, 4.722422410127076e-06)
item 원예용품 has different mean with total dataset; (3.3055864179545322, 0.000980404693903099)
item 유가공 has different mean with total dataset; (8.474975462762382, 7.804558763088175e-17)
item 유아식기 has different mean with total dataset; (3.2586277178437193, 0.0011482943806549887)
item 유아용품 has different mean with total dataset; (3.084334522837267, 0.002093171459059815)
item 육포 has different mean with total dataset; (5.701571035526521, 1.5416521623717162e-08)
item 음반영상 has different mean with total dataset; (2.4111176557170233, 0.016070592963181283)
item 이온음료 has different mean with total dataset; (6.5475724456555895, 9.176703702146386e-11)
item 이유식/유아간식 has different mean with total dataset; (5.5333976388521, 3.996838473014277e-08)
item 이지캐주얼브랜드 has different mean with total dataset; (13.304128905965118, 2.341667002069992e-37)
item 인스턴트선물세트 has different mean with total dataset; (2.6244645567410427, 0.008799853232568856)
item 일반욕실용품 has different mean with total dataset; (8.568983195164765, 3.56255795503135e-17)
item 일반헤어케어 has different mean with total dataset; (5.752306204148443, 1.156614841735934e-08)
item 자동차방향제/소취제 has different mean with total dataset; (5.759569684958471, 1.1038672984159629e-08)
item 자몽 has different mean with total dataset; (6.35641367452082, 3.077803405507466e-10)
item 잡화멀티샵 has different mean with total dataset; (3.339027342183163, 0.0008706467328397145)
item 장식용품 has different mean with total dataset; (2.701437211425631, 0.007012684433664166)
item 전통주 has different mean with total dataset; (12.206479467574448, 4.475394493930141e-32)
item 제습제 has different mean with total dataset; (4.639521498976986, 3.93716246247134e-06)
item 조기/명태 has different mean with total dataset; (5.004545817165185, 6.578649968708504e-07)
item 조리반찬 has different mean with total dataset; (6.10688343497605, 1.4340516079078073e-09)
item 조명기구 has different mean with total dataset; (3.367321612847052, 0.000785670877062855)
item 주니어양말 has different mean with total dataset; (2.9358670495445724, 0.0033979724422648905)
item 준보석 has different mean with total dataset; (3.888234574377179, 0.00010733507121506645)
item 즉석구이김 has different mean with total dataset; (4.378971965702216, 1.3117886891385495e-05)
item 즉석구이안주 has different mean with total dataset; (3.6861321873398154, 0.00023893607234442445)
item 즉석도정미 has different mean with total dataset; (4.594323132519724, 4.864887731845669e-06)
item 차 has different mean with total dataset; (7.618326112057484, 5.6590122880599855e-14)
item 천연조미료 has different mean with total dataset; (2.9515694520376714, 0.0032272697070919506)
item 캐주얼슈즈 has different mean with total dataset; (2.365537012943426, 0.01817790948681538)
item 캠핑 has different mean with total dataset; (2.7848206598219964, 0.005451281817607978)
item 커피선물세트 has different mean with total dataset; (2.328916913422505, 0.020047424352357863)
item 튀김 has different mean with total dataset; (8.677521363361915, 1.5523559638988482e-17)
item 특산물행사 has different mean with total dataset; (5.330150794897585, 1.2024150037460052e-07)
item 파인애플 has different mean with total dataset; (3.8246762103927754, 0.00013885426018664732)
item 편의시설 has different mean with total dataset; (2.619092043934406, 0.008930957265780113)
item 포장반찬HMR has different mean with total dataset; (4.2350867878887835, 2.4882732841116502e-05)
item 포장젓갈 has different mean with total dataset; (4.6852745760784265, 3.1689252920008937e-06)
item 풋/핸드케어 has different mean with total dataset; (4.443030668154154, 9.820205546586495e-06)
item 피자 has different mean with total dataset; (2.50942202117538, 0.012230322183084806)
item 한방차 has different mean with total dataset; (3.9766576141812755, 7.472808441754368e-05)
item 해초 has different mean with total dataset; (7.016562485764557, 4.080312601371517e-12)
item 헬스용품 has different mean with total dataset; (5.413637262463601, 7.66523935179359e-08)
item 호주산소고기 has different mean with total dataset; (13.714370531980986, 2.0604862076459165e-39)
item 홍인삼 has different mean with total dataset; (3.5666295811266493, 0.00037787326937315164)
item 황태 has different mean with total dataset; (7.44352609383247, 2.0520382897497898e-13)
item 훼이셜케어 has different mean with total dataset; (6.09496252220006, 1.5413571802284243e-09)

for cluster 11:

item 가공식품 has different mean with total dataset; (11.267338761139396, 1.230706902311849e-27)
item 가구 has different mean with total dataset; (3.632329661082308, 0.00029609669396605094)
item 골프 has different mean with total dataset; (2.4472067704660314, 0.014602399630841917)
item 골프용품 has different mean with total dataset; (14.94293797836554, 7.407873347928993e-45)
item 과자 has different mean with total dataset; (9.1888540915786, 3.1690840036893575e-19)
item 남성정장 has different mean with total dataset; (6.720387110226764, 3.315994736775047e-11)
item 남성케어 has different mean with total dataset; (3.529517774810427, 0.00043833169375004144)
item 대형가전 has different mean with total dataset; (4.593724940316933, 5.002726544462383e-06)
item 모피/피혁 has different mean with total dataset; (3.1208639462353025, 0.0018640003484219861)
item 보석 has different mean with total dataset; (2.9260544218830993, 0.0035219665872954294)
item 선케어 has different mean with total dataset; (2.9861656422897718, 0.002906814495019906)
item 시티웨어 has different mean with total dataset; (3.0344190275976417, 0.002470498193023191)
item 아이메이크업 has different mean with total dataset; (2.845524272125776, 0.004540683320669033)
item 준보석/시계 has different mean with total dataset; (3.8323436503275916, 0.00013590860107869792)
item 피혁잡화 has different mean with total dataset; (2.529510261849613, 0.011509535393516882)

for cluster 12:

item 레스토랑 has different mean with total dataset; (3.7162288974480937, 0.00020895792944117608)
item 카페 has different mean with total dataset; (2.018117260822943, 0.04373318236422914)
item 컨세션 has different mean with total dataset; (3.4374066775338084, 0.0006023751725724242)

for cluster 13:

item 피혁잡화 has different mean with total dataset; (1.9822767241395405, 0.04753098993527463)

남성 사용자를 클러스터링해 보았음. 이 때 6번 클러스터와 7번 클러스터 간의 차이를 파악하였음.

In [31]:
pt_male= pt_bytsne.loc[df_demo['IS_FEMALE']==0, :].reset_index(drop= True)
df_male= df_item_consumed.loc[df_demo['IS_FEMALE']==0, :].drop('ID', axis= 1).reset_index(drop= True)
id_male= df_raw.loc[df_raw['IS_MALE']==1, 'ID'].drop_duplicates().sort_values().reset_index(drop= True)
df_demo_male= df_demo.loc[df_demo['ID'].isin(id_male), :].reset_index(drop= True)
df_male.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 7003 entries, 0 to 7002
Columns: 149 entries, H&B선물세트 to 훼이셜케어
dtypes: float64(149)
memory usage: 8.0 MB
In [32]:
gmm = GaussianMixture(n_components=14, covariance_type='full', random_state= 42).fit(pt_male[[0,1]])
label_dict[(2, 14)]= pd.Series(gmm.predict(pt_male[[0,1]]))
plot_results(pt_male[[0,1]], label_dict[(2,14)])
In [33]:
plot_clusters(pt_male[[0,1]], label_dict[(2,14)])
In [34]:
find_prefered_item(df_male, label_dict[(2,14)])

for cluster 0:

item 탈취제 has different mean with total dataset; (2.0419680352992704, 0.04160079302594415)
item 훼이셜케어 has different mean with total dataset; (2.3323599388324268, 0.02004032386316317)

for cluster 1:

item 가공식품 has different mean with total dataset; (2.913083654507895, 0.003709770570241837)
item 가공우유 has different mean with total dataset; (15.67551936206713, 1.0153405456487646e-51)
item 봉투보증금 has different mean with total dataset; (3.9168124709859797, 0.00010124557716052506)

for cluster 2:

item 가공식품 has different mean with total dataset; (10.704918560775756, 1.373045410334679e-23)
item 골프용품 has different mean with total dataset; (14.321693222971092, 2.474838088644525e-37)
item 과자 has different mean with total dataset; (5.273221169296936, 2.2872819634808253e-07)
item 남성정장 has different mean with total dataset; (6.882567205683728, 2.4693290119539018e-11)
item 남성케어 has different mean with total dataset; (2.552784613110995, 0.011082423476798399)
item 대형가전 has different mean with total dataset; (2.5877869588650806, 0.010000996427962334)
item 모피/피혁 has different mean with total dataset; (2.0454275781784377, 0.04150714778266778)
item 보석 has different mean with total dataset; (2.2233062782998236, 0.026759633374186305)
item 시티웨어 has different mean with total dataset; (4.075568854101397, 5.5938110810478806e-05)
item 아이메이크업 has different mean with total dataset; (2.6214154748487926, 0.009108741973532503)
item 피혁잡화 has different mean with total dataset; (4.789220370093036, 2.406908180115132e-06)

for cluster 3:

item 구기용품 has different mean with total dataset; (2.972747175866649, 0.003110008802776471)
item 김선물세트 has different mean with total dataset; (2.0351523302097454, 0.04242394947176853)
item 라켓스포츠 has different mean with total dataset; (2.2678383172248053, 0.023816462579218173)
item 레스토랑 has different mean with total dataset; (4.885639950122681, 1.4196042678041805e-06)
item 매트 has different mean with total dataset; (2.254397077553506, 0.024648533479233275)
item 문화센터 has different mean with total dataset; (3.2990940332934153, 0.001046171809604432)
item 미시캐주얼 has different mean with total dataset; (2.147990191318627, 0.032250699608419336)
item 복합샵 has different mean with total dataset; (4.694311555500335, 3.5806523029762396e-06)
item 브랜드침구 has different mean with total dataset; (2.3169118028309055, 0.02095371178631901)
item 수예시즌 has different mean with total dataset; (2.8902003053327308, 0.004034914376988583)
item 수족관용품 has different mean with total dataset; (4.241883486466626, 2.7061853513791318e-05)
item 스낵 has different mean with total dataset; (3.8027211368284823, 0.00016129541566598836)
item 스타킹 has different mean with total dataset; (2.111840788632194, 0.035230959512874624)
item 시즌슈즈 has different mean with total dataset; (3.2554095069805906, 0.0012170826126511602)
item 시즌잡화 has different mean with total dataset; (2.206985158815745, 0.027813662661804697)
item 시즌행사 has different mean with total dataset; (3.334271719830449, 0.0009271235930525333)
item 시티캐주얼 has different mean with total dataset; (1.9714217340847513, 0.04930435407350708)
item 아동브랜드 has different mean with total dataset; (7.3810797319621075, 7.966858864245484e-13)
item 아동침구 has different mean with total dataset; (2.2780943150745006, 0.023185471507279207)
item 양말선물세트 has different mean with total dataset; (2.005679353808146, 0.04547892307434956)
item 언더웨어행사 has different mean with total dataset; (2.3095935335712707, 0.021360831995792872)
item 여성양말 has different mean with total dataset; (3.6287450440898237, 0.0003169582750408797)
item 여성캐주얼브랜드 has different mean with total dataset; (3.0386852512007105, 0.0025180881669826555)
item 우산/양산 has different mean with total dataset; (2.887114348186227, 0.0040691959294507545)
item 윈터스포츠 has different mean with total dataset; (2.198367240845784, 0.02844105205579637)
item 유아식기 has different mean with total dataset; (4.139464682619142, 4.166247807000225e-05)
item 유아토들러 has different mean with total dataset; (3.379091824135399, 0.0007909530125333136)
item 이온음료 has different mean with total dataset; (2.6685434360023033, 0.007856134673208602)
item 이유식/유아간식 has different mean with total dataset; (5.132637155066544, 4.308807503177262e-07)
item 일반욕실용품 has different mean with total dataset; (5.221054165070769, 2.719850541710847e-07)
item 일반헤어케어 has different mean with total dataset; (3.5257120479883386, 0.0004643100336796955)
item 자동차방향제/소취제 has different mean with total dataset; (3.510092892659699, 0.0004917085947712009)
item 자전거 has different mean with total dataset; (2.554575990288454, 0.010960057709559006)
item 장갑 has different mean with total dataset; (2.7161843480482433, 0.0068595924837366)
item 장착용품 has different mean with total dataset; (2.6159317780145117, 0.0091935616519526)
item 즉석구이안주 has different mean with total dataset; (3.0291362594800573, 0.0025935419397651333)
item 차 has different mean with total dataset; (3.995939742771929, 7.46207660684358e-05)
item 캐주얼화브랜드 has different mean with total dataset; (2.6168272967904813, 0.009179947055307893)
item 캠핑 has different mean with total dataset; (2.545896637245415, 0.011235085262760128)
item 탈취제 has different mean with total dataset; (3.8170056712681006, 0.00015336883099826215)
item 튀김 has different mean with total dataset; (2.5567107560715265, 0.01085271867333021)
item 파인애플 has different mean with total dataset; (2.1372454959861376, 0.03305292774917084)
item 판매시설 has different mean with total dataset; (3.957595317211306, 8.825338672015047e-05)
item 편의시설 has different mean with total dataset; (3.894264284616233, 0.00011284386796618654)
item 푸드코트 has different mean with total dataset; (6.358787058538393, 5.130273628724819e-10)
item 해초 has different mean with total dataset; (2.0961049144393082, 0.03662544597324381)
item 헬스용품 has different mean with total dataset; (2.956346831633282, 0.003274880325784192)
item 황태 has different mean with total dataset; (2.3575736588904985, 0.01880850053065653)

for cluster 4:

item 가구 has different mean with total dataset; (2.5531114773749217, 0.011126109506570154)
item 보석 has different mean with total dataset; (2.6042369003875896, 0.009653176618250757)
item 아이메이크업 has different mean with total dataset; (2.840116133137027, 0.004816887546629855)
item 준보석/시계 has different mean with total dataset; (3.1420885010251975, 0.0017970920102607236)
item 피혁잡화 has different mean with total dataset; (2.4231047066744353, 0.015836499944358564)

for cluster 5:

item 브랜드침구 has different mean with total dataset; (2.1258597652155395, 0.03389321107935852)
item 수족관용품 has different mean with total dataset; (3.47532518711405, 0.000540651927854437)
item 유아식기 has different mean with total dataset; (2.416509305700306, 0.015939803116515463)
item 유아토들러 has different mean with total dataset; (2.35277823410631, 0.01892634349411641)
item 자동차방향제/소취제 has different mean with total dataset; (3.7027826733822584, 0.0002310246470889382)
item 편의시설 has different mean with total dataset; (5.006558311379343, 7.175157626086333e-07)

for cluster 6:

item H&B선물세트 has different mean with total dataset; (2.8645800250485838, 0.004272883969154707)
item 가공식품 has different mean with total dataset; (2.9130202119802306, 0.0036573220980147263)
item 가방브랜드 has different mean with total dataset; (2.865239107974792, 0.004270826044619828)
item 골프 has different mean with total dataset; (2.5080874288924195, 0.012310421289269381)
item 기능성건강 has different mean with total dataset; (4.530929332339479, 6.690228587958173e-06)
item 김선물세트 has different mean with total dataset; (2.68127878908034, 0.007474499978014363)
item 남성어덜트캐주얼 has different mean with total dataset; (2.0253036190017912, 0.043147375956981336)
item 드레스화브랜드 has different mean with total dataset; (3.4605310545332197, 0.0005657836265917922)
item 라켓스포츠 has different mean with total dataset; (3.1354733016310536, 0.0017727352551844212)
item 레스토랑 has different mean with total dataset; (3.711724190928449, 0.0002180651336918613)
item 벽지류 has different mean with total dataset; (2.959043595026943, 0.0031686505526249444)
item 브랜드침구 has different mean with total dataset; (2.19104557573276, 0.028696708328690142)
item 비즈니스정장 has different mean with total dataset; (2.965346253655459, 0.003103244201814422)
item 석류 has different mean with total dataset; (3.7694575212433112, 0.00017491037167704165)
item 소가구 has different mean with total dataset; (2.0875029684379527, 0.03713234202165671)
item 수예시즌 has different mean with total dataset; (2.801195394081721, 0.005205044554201556)
item 수입게 has different mean with total dataset; (2.65243840589036, 0.008137712381580798)
item 수족관용품 has different mean with total dataset; (1.969343120582632, 0.04918863984735658)
item 스낵 has different mean with total dataset; (15.07923787388969, 1.108620755593669e-45)
item 스타킹 has different mean with total dataset; (4.524137013654239, 6.926175582433539e-06)
item 스포츠ACC has different mean with total dataset; (2.6955812438715903, 0.007162530981896541)
item 스포츠슈즈 has different mean with total dataset; (2.7419054581814843, 0.006236777511745537)
item 시계 has different mean with total dataset; (2.832435995779153, 0.004731184688057315)
item 시즌슈즈 has different mean with total dataset; (2.0960860480062165, 0.036357571664495915)
item 시즌잡화 has different mean with total dataset; (2.20211113153408, 0.027915379978734556)
item 시즌행사 has different mean with total dataset; (5.031709164569451, 5.83812707730818e-07)
item 시티캐주얼 has different mean with total dataset; (3.841510316337707, 0.00012983381830016595)
item 식사류 has different mean with total dataset; (4.5365793171348665, 6.567365255774932e-06)
item 식자재 has different mean with total dataset; (2.002240032953679, 0.045580708061196074)
item 아동브랜드 has different mean with total dataset; (2.600234874026553, 0.009442276428687814)
item 아로마 has different mean with total dataset; (3.0615985587200307, 0.0022669327049412145)
item 애견용품 has different mean with total dataset; (7.395603690605324, 3.465284854742817e-13)
item 언더웨어행사 has different mean with total dataset; (3.586937639050662, 0.0003531972502871889)
item 여성양말 has different mean with total dataset; (3.572212193852571, 0.00037399697271913574)
item 여성캐주얼브랜드 has different mean with total dataset; (3.412037053165797, 0.0006721019782997301)
item 우산/양산 has different mean with total dataset; (1.995218598491354, 0.04633775314333988)
item 원예용품 has different mean with total dataset; (3.830427493377676, 0.00013679508968591166)
item 유가공 has different mean with total dataset; (5.879772533795735, 5.714501390902904e-09)
item 유아식기 has different mean with total dataset; (2.3894266397162918, 0.017065133715184252)
item 유아용품 has different mean with total dataset; (2.767059786807282, 0.005761092357536738)
item 육포 has different mean with total dataset; (2.8820858453798057, 0.004052515388012787)
item 이온음료 has different mean with total dataset; (2.9479764508890764, 0.0032889179587444624)
item 이지캐주얼브랜드 has different mean with total dataset; (11.638763275804132, 4.129094116556643e-29)
item 인스턴트선물세트 has different mean with total dataset; (2.0189011552099956, 0.0437661770953411)
item 일반욕실용품 has different mean with total dataset; (6.75005685821791, 2.6258622771840757e-11)
item 일반헤어케어 has different mean with total dataset; (4.9417602579447175, 9.295584404761301e-07)
item 자동차방향제/소취제 has different mean with total dataset; (3.6297614178887283, 0.00029942016675233964)
item 자몽 has different mean with total dataset; (3.6921492138072733, 0.00023653612262687447)
item 잡화멀티샵 has different mean with total dataset; (2.9737211658681386, 0.003020805188820906)
item 장갑 has different mean with total dataset; (2.213016389358471, 0.027151736415469155)
item 전통주 has different mean with total dataset; (7.958991679265523, 5.725595121078969e-15)
item 제습제 has different mean with total dataset; (4.119148120266279, 4.1746430585906784e-05)
item 조기/명태 has different mean with total dataset; (3.127920149046755, 0.0018066056765465477)
item 조리반찬 has different mean with total dataset; (7.554280171655837, 1.1017746418038177e-13)
item 조명기구 has different mean with total dataset; (3.83004731662602, 0.00013743060953763615)
item 준보석 has different mean with total dataset; (1.9714853689124148, 0.048978481623787526)
item 즉석구이김 has different mean with total dataset; (3.65659344485178, 0.00027108580266699327)
item 즉석구이안주 has different mean with total dataset; (2.766513791551431, 0.005775116743191034)
item 즉석도정미 has different mean with total dataset; (5.237706500602289, 2.0469226561120545e-07)
item 차 has different mean with total dataset; (5.85888667126467, 6.222776301146665e-09)
item 천연조미료 has different mean with total dataset; (2.9885051677582117, 0.002878744469735131)
item 캐주얼슈즈 has different mean with total dataset; (2.42831818265167, 0.015377539297081357)
item 커튼류 has different mean with total dataset; (2.169105856686602, 0.03032220944450543)
item 탈취제 has different mean with total dataset; (5.752722849180387, 1.220806531605932e-08)
item 튀김 has different mean with total dataset; (5.875977602252177, 6.076469967589694e-09)
item 특산물행사 has different mean with total dataset; (2.314088850524821, 0.020907069495136675)
item 파인애플 has different mean with total dataset; (2.8291984510749426, 0.0047782477485029924)
item 편의시설 has different mean with total dataset; (1.9705112276812438, 0.04901157906704022)
item 포장반찬HMR has different mean with total dataset; (2.8969635652336234, 0.003863119623079686)
item 포장젓갈 has different mean with total dataset; (2.8626939291932882, 0.004304647996755682)
item 풋/핸드케어 has different mean with total dataset; (3.1582101668007243, 0.0016422688213157273)
item 한방차 has different mean with total dataset; (3.26330887734132, 0.0011436212729609636)
item 해초 has different mean with total dataset; (5.1729402412928716, 2.8783522312129575e-07)
item 헬스용품 has different mean with total dataset; (2.3833549804313, 0.017359771125115497)
item 호주산소고기 has different mean with total dataset; (4.900551394267607, 1.15106486813265e-06)
item 홍인삼 has different mean with total dataset; (3.0052427122821657, 0.002732421127716601)
item 황태 has different mean with total dataset; (6.930782541515521, 8.228779591724916e-12)
item 훼이셜케어 has different mean with total dataset; (5.537376661714225, 4.078354700682659e-08)

for cluster 7:

item 가공식품 has different mean with total dataset; (8.124185006598912, 6.597258281620677e-15)
item 가공우유 has different mean with total dataset; (19.211569033581895, 9.159328149272303e-57)
item 계절완구 has different mean with total dataset; (2.1911508034657996, 0.029092502259485915)
item 고급 has different mean with total dataset; (2.255805516844525, 0.024697990972152466)
item 과일야채 has different mean with total dataset; (4.664772573196922, 4.389235587592322e-06)
item 너트류 has different mean with total dataset; (3.5172101687250374, 0.0004913876259434618)
item 레스토랑 has different mean with total dataset; (3.094887695897716, 0.0021221831077342553)
item 롤샌드/핫도그 has different mean with total dataset; (2.692064891692918, 0.007438561530149789)
item 봉투보증금 has different mean with total dataset; (2.8439408815413443, 0.004717468245232697)
item 샐러드 has different mean with total dataset; (2.2957567457562456, 0.022280377589411617)
item 스낵 has different mean with total dataset; (4.834300380374094, 1.982573950002326e-06)
item 스타킹 has different mean with total dataset; (3.3047402826461987, 0.0010474833709862212)
item 시즌행사 has different mean with total dataset; (2.8307071599012787, 0.004881865510621555)
item 식빵 has different mean with total dataset; (2.880809820506915, 0.004204512718743634)
item 아동브랜드 has different mean with total dataset; (2.17166409862103, 0.03050838882281106)
item 아동완구 has different mean with total dataset; (3.9335185595453463, 0.00010075594008362078)
item 여성양말 has different mean with total dataset; (2.048202474561283, 0.04124638207912973)
item 오뎅 has different mean with total dataset; (2.9287350841522373, 0.0036224540252915494)
item 유가공 has different mean with total dataset; (3.061710085037312, 0.002358988143520166)
item 유아식기 has different mean with total dataset; (3.4129655292715904, 0.0007147737648513968)
item 유아토들러 has different mean with total dataset; (2.0318466118306215, 0.042896580916949034)
item 육포 has different mean with total dataset; (3.3756319966823503, 0.0008153101935765778)
item 이온음료 has different mean with total dataset; (3.263311015778591, 0.0011956945747230204)
item 일반욕실용품 has different mean with total dataset; (3.326708281358391, 0.0009696538007451824)
item 일반헤어케어 has different mean with total dataset; (2.055937064466877, 0.040464801284428385)
item 자동차방향제/소취제 has different mean with total dataset; (2.465636220468077, 0.014133812001705118)
item 제습제 has different mean with total dataset; (2.509314876004254, 0.012533791411349741)
item 조리면 has different mean with total dataset; (1.9991013854706157, 0.046336256062162365)
item 즉석구이안주 has different mean with total dataset; (2.4839540002535268, 0.01344754980975949)
item 차 has different mean with total dataset; (2.0028899047623177, 0.04595780937810826)
item 튀김 has different mean with total dataset; (2.349853781020804, 0.019294578817331844)
item 특수점완구 has different mean with total dataset; (2.561272549951745, 0.010846965599788646)
item 파인애플 has different mean with total dataset; (2.612767930042751, 0.00935732996320652)
item 편의시설 has different mean with total dataset; (3.4513796361785656, 0.0006212746721018234)
item 편의점카페 has different mean with total dataset; (3.6648910640869548, 0.0002854155021027574)
item 프리미엄 has different mean with total dataset; (2.656759899770461, 0.008250828033108495)
item 피자 has different mean with total dataset; (2.177544564753559, 0.03006514682369068)
item 피혁잡화 has different mean with total dataset; (2.490118952571643, 0.0132048883825917)
item 필기용품 has different mean with total dataset; (2.7356211790929135, 0.0065418962753370485)
item 호주산소고기 has different mean with total dataset; (2.366829390448908, 0.018342063591211696)
item 화과자 has different mean with total dataset; (4.68775658208455, 3.9590573349151085e-06)
item 황태 has different mean with total dataset; (2.7734340019045423, 0.005835869025695784)

for cluster 8:

item 고양이용품 has different mean with total dataset; (2.784013836001479, 0.005480730736405332)
item 기능성건강 has different mean with total dataset; (2.6479319271352146, 0.008232544085116264)
item 레스토랑 has different mean with total dataset; (3.0239197865546985, 0.00256327322179412)
item 스낵 has different mean with total dataset; (4.018476307247489, 6.128362332960945e-05)
item 시티캐주얼 has different mean with total dataset; (3.7858816125610515, 0.00016198148534657094)
item 여성캐주얼브랜드 has different mean with total dataset; (2.186753287513189, 0.02897110077965544)
item 유아용품 has different mean with total dataset; (3.1226045172090044, 0.00184916187921272)
item 일반헤어케어 has different mean with total dataset; (2.2792073570801565, 0.02286385500667511)
item 자몽 has different mean with total dataset; (2.1902177248790307, 0.02874627529474243)
item 차 has different mean with total dataset; (2.6417537177590176, 0.008370095594126995)
item 튀김 has different mean with total dataset; (7.864295226420433, 8.610441467994911e-15)
item 피자 has different mean with total dataset; (2.6481992536971513, 0.008229719576294935)

for cluster 9:


for cluster 10:

item 골프용품 has different mean with total dataset; (5.364397805375737, 1.0109384395342842e-07)
item 남성정장 has different mean with total dataset; (6.931738303901187, 1.2026648168810738e-11)
item 시티웨어 has different mean with total dataset; (6.603368654166181, 9.568697426503048e-11)
item 피혁잡화 has different mean with total dataset; (10.254218381494848, 1.160777653095967e-22)

for cluster 11:

item 주유소 has different mean with total dataset; (5.19868293618637, 4.154161480839718e-07)

for cluster 12:

item 가공식품 has different mean with total dataset; (15.092228162909345, 4.912676697378037e-41)
item 고급 has different mean with total dataset; (2.977117334037037, 0.00308841604567082)
item 남성정장 has different mean with total dataset; (2.216866337716037, 0.027165227748499308)
item 대형가전 has different mean with total dataset; (3.5447178987372214, 0.00044048799927398505)
item 보석 has different mean with total dataset; (2.32338183039912, 0.020668777864757723)
item 카페 has different mean with total dataset; (4.3597143968927785, 1.6654020736884933e-05)
item 컨세션 has different mean with total dataset; (3.992165963899153, 7.817003085745691e-05)
item 피혁잡화 has different mean with total dataset; (4.08693515967627, 5.250488713182798e-05)

for cluster 13:
In [35]:
find_prefered_by_clusters(df_male, label_dict[(2, 14)], 6, 7)

item 가공식품 has different mean with total dataset; (-5.832776731086528, 9.253239253375856e-09)
item 가공우유 has different mean with total dataset; (-19.444665082544372, 8.563636933160878e-58)
item 가방브랜드 has different mean with total dataset; (3.194662174154414, 0.0014448165350527367)
item 계절완구 has different mean with total dataset; (-2.4690126769472807, 0.01401974842473204)
item 고급 has different mean with total dataset; (-2.415875717541854, 0.016207801012808824)
item 과일야채 has different mean with total dataset; (-4.4200640098135855, 1.2666329490414027e-05)
item 남성어덜트캐주얼 has different mean with total dataset; (2.1638115635387876, 0.030692726084018334)
item 너트류 has different mean with total dataset; (-3.872620362067213, 0.00012820677393496233)
item 동물병원 has different mean with total dataset; (2.093468473211416, 0.036618722349524485)
item 드레스화브랜드 has different mean with total dataset; (2.328347974089112, 0.02008021487621238)
item 롤샌드/핫도그 has different mean with total dataset; (-2.639488874001812, 0.008637120566239467)
item 봉투보증금 has different mean with total dataset; (-2.784919575761394, 0.005638414157322959)
item 브랜드침구 has different mean with total dataset; (2.2863239564863225, 0.022418944574227676)
item 샐러드 has different mean with total dataset; (-2.436825795583181, 0.015316366415564038)
item 석류 has different mean with total dataset; (2.222682637476845, 0.02646831087963648)
item 수예시즌 has different mean with total dataset; (2.030388280832931, 0.04255410275956088)
item 스낵 has different mean with total dataset; (5.555522778880374, 3.806377686764849e-08)
item 스포츠슈즈 has different mean with total dataset; (2.292218472472571, 0.022073645555220078)
item 시계 has different mean with total dataset; (2.4353361975332284, 0.015036272646873487)
item 시티캐주얼 has different mean with total dataset; (3.267782086244287, 0.001121588826550323)
item 식빵 has different mean with total dataset; (-2.9835378872555918, 0.0030415662273600326)
item 식사류 has different mean with total dataset; (3.4829644754204656, 0.0005152084196098118)
item 식자재 has different mean with total dataset; (2.639158218058829, 0.008472945937442725)
item 아동완구 has different mean with total dataset; (-4.354748194326285, 1.7536605691764328e-05)
item 애견용품 has different mean with total dataset; (6.996924195643474, 4.686515122546557e-12)
item 양말선물세트 has different mean with total dataset; (2.195042672335063, 0.028363537565326)
item 유아식기 has different mean with total dataset; (-2.0202261799791863, 0.043875386551594134)
item 이지캐주얼브랜드 has different mean with total dataset; (10.930949575907317, 2.1748300063178412e-26)
item 일반헤어케어 has different mean with total dataset; (2.351845309337989, 0.018864692912648404)
item 잡화멀티샵 has different mean with total dataset; (2.1428951635538493, 0.03235320633641319)
item 전통주 has different mean with total dataset; (7.816508653210784, 1.4787390052912232e-14)
item 조리면 has different mean with total dataset; (-2.274481983049914, 0.02353623899154671)
item 조리반찬 has different mean with total dataset; (2.56962086864148, 0.010403420017907014)
item 조명기구 has different mean with total dataset; (2.3922374409523233, 0.016938336706287)
item 준보석 has different mean with total dataset; (3.2998316455372216, 0.0010102423489436268)
item 준보석/시계 has different mean with total dataset; (-2.2347724799240334, 0.02604189169425219)
item 즉석도정미 has different mean with total dataset; (2.313150683852437, 0.0209531200834425)
item 탈취제 has different mean with total dataset; (3.4360451389384887, 0.000611905845386143)
item 튀김 has different mean with total dataset; (3.5188806964769768, 0.0004503631405326504)
item 특산물행사 has different mean with total dataset; (2.01864215083147, 0.04379148165542096)
item 특수점완구 has different mean with total dataset; (-2.6803272838126895, 0.007700195953116787)
item 판매용소모품 has different mean with total dataset; (-2.0291942998036157, 0.04319338839249055)
item 편의시설 has different mean with total dataset; (-2.4169452724262497, 0.016030182659162083)
item 편의점카페 has different mean with total dataset; (-3.6463973483775454, 0.0003052260057839738)
item 포장반찬HMR has different mean with total dataset; (2.7838633190095865, 0.005460504985357176)
item 프리미엄 has different mean with total dataset; (-2.7556364637367055, 0.0061643070821951786)
item 피혁잡화 has different mean with total dataset; (-2.177754406231949, 0.02988052629628526)
item 필기용품 has different mean with total dataset; (-2.5787961061361715, 0.010289196648537781)
item 해초 has different mean with total dataset; (2.444173644408155, 0.014700209457932413)
item 호주산소고기 has different mean with total dataset; (4.022638145182845, 6.223954793628303e-05)
item 화과자 has different mean with total dataset; (-4.892105940098577, 1.5106426771677516e-06)
item 훼이셜케어 has different mean with total dataset; (2.6730814208017932, 0.007651717501852715)
In [36]:
df_byitem.loc[df_demo['IS_MALE']==1].reset_index().loc[label_dict[(2,14)]==6, :].head()
Out[36]:
index ID H&B선물세트 VIDEOGAME 가공식품 가공우유 가구 가방브랜드 계절완구 고급 ... 필기용품 한방차 한우선물세트 해초 헬스용품 호주산소고기 홍인삼 화과자 황태 훼이셜케어
1 3 4 0.0 0.0 38.0 0.0 0.0 0.0 0.0 0.0 ... 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
12 26 29 0.0 0.0 11.0 0.0 0.0 0.0 0.0 0.0 ... 0.0 0.0 0.0 0.0 0.0 9.0 0.0 0.0 4.0 1.0
19 39 42 0.0 0.0 22.0 0.0 0.0 0.0 0.0 0.0 ... 0.0 3.0 0.0 0.0 3.0 0.0 0.0 0.0 1.0 1.0
22 46 50 0.0 0.0 9.0 2.0 0.0 0.0 0.0 0.0 ... 0.0 1.0 0.0 1.0 0.0 15.0 14.0 0.0 0.0 0.0
26 52 56 0.0 0.0 13.0 0.0 0.0 0.0 0.0 0.0 ... 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0

5 rows × 151 columns

6번 클러스터와 7번 클러스터 간의 지표 차이를 파악하였음.

In [37]:
df_clst6= df_demo_male.loc[label_dict[(2,14)]==6, :]
In [38]:
df_clst7= df_demo_male.loc[label_dict[(2,14)]==7, :]
In [39]:
df_info6= df_grouped.loc[df_grouped['ID'].isin(df_clst6['ID']), :]
df_info7= df_grouped.loc[df_grouped['ID'].isin(df_clst7['ID']), :]
In [40]:
df_info6.head()
Out[40]:
ID RCT_NO BIZ_UNIT PD_S_C BR_C BUY_AM
3 4 308 3 136 7 7970116
26 29 408 2 159 5 3756730
39 42 492 2 138 3 5425186
46 50 566 3 238 11 10626666
52 56 107 3 58 6 2197496
In [41]:
df_clst6.AGE_PRD.value_counts().sort_index()
Out[41]:
20     55
30    156
40    249
50    212
60    129
Name: AGE_PRD, dtype: int64
In [42]:
df_value6= df_clst6.PRVIN.value_counts().sort_values()
df_value7= df_clst7.PRVIN.value_counts().sort_values()
df_ratio6= (df_value6/df_clst6.PRVIN.count()).apply(lambda x: round(x, 3))
df_ratio7= (df_value7/df_clst7.PRVIN.count()).apply(lambda x: round(x, 3))

num_of_range= range(1, df_value6.count()+1)
grp= plt.figure(figsize= (9,12))
ax= grp.add_subplot(211)
plt.hlines(y=num_of_range, xmin=0, xmax=df_ratio6, color='skyblue')
for i in num_of_range:
    ax.annotate(df_ratio6.values.tolist()[i-1], xy=(df_ratio6.values.tolist()[i-1]+0.01, i+0.01), textcoords='data')
plt.plot(df_ratio6, num_of_range, "o")
plt.yticks(num_of_range, df_value6.index)

num_of_range= range(1, df_value7.count()+1)
grp= plt.figure(figsize= (9,9))
ax= grp.add_subplot(212)
plt.hlines(y=num_of_range, xmin=0, xmax=df_ratio7, color='tomato')
for i in num_of_range:
    ax.annotate(df_ratio7.values.tolist()[i-1], xy=(df_ratio7.values.tolist()[i-1]+0.01, i+0.01), textcoords='data')
plt.plot(df_ratio7, num_of_range, "o", color= 'red')
plt.yticks(num_of_range, df_value6.index)
plt.show()
In [43]:
df_value6= df_clst6.AGE_PRD.value_counts().sort_index()
df_value7= df_clst7.AGE_PRD.value_counts().sort_index()
df_ratio6= (df_value6/df_clst6.AGE_PRD.count()).apply(lambda x: round(x, 3))
df_ratio7= (df_value7/df_clst7.AGE_PRD.count()).apply(lambda x: round(x, 3))

num_of_range= range(1, df_value6.count()+1)
grp= plt.figure(figsize= (9,12))
ax= grp.add_subplot(211)
plt.hlines(y=num_of_range, xmin=0, xmax=df_ratio6, color='skyblue')
for i in num_of_range:
    ax.annotate(df_ratio6.values.tolist()[i-1], xy=(df_ratio6.values.tolist()[i-1]+0.01, i+0.01), textcoords='data')
plt.plot(df_ratio6, num_of_range, "o")
plt.yticks(num_of_range, df_value6.index)

num_of_range= range(1, df_value7.count()+1)
grp= plt.figure(figsize= (9,9))
ax= grp.add_subplot(212)
plt.hlines(y=num_of_range, xmin=0, xmax=df_ratio7, color='tomato')
for i in num_of_range:
    ax.annotate(df_ratio7.values.tolist()[i-1], xy=(df_ratio7.values.tolist()[i-1]+0.01, i+0.01), textcoords='data')
plt.plot(df_ratio7, num_of_range, "o", color= 'red')
plt.yticks(num_of_range, df_value6.index)
plt.show()
In [44]:
grp= sns.distplot(df_info6['BUY_AM'], label= 'cluster 6', bins= 30, kde= False, norm_hist= True)
grp= sns.distplot(df_info7['BUY_AM'], label= 'cluster 7', kde= False, norm_hist= True)
plt.legend()
plt.show()
In [45]:
grp= sns.distplot(df_info6['RCT_NO'], label= 'cluster 6', bins= 30, kde= False, norm_hist= True)
grp= sns.distplot(df_info7['RCT_NO'], label= 'cluster 7', kde= False, norm_hist= True)
plt.legend()
plt.show()
In [46]:
grp= sns.distplot(df_info6['PD_S_C'], label= 'cluster 6', bins= 30, kde= True)
grp= sns.distplot(df_info7['PD_S_C'], label= 'cluster 7', kde= True)
plt.legend()
plt.show()
In [47]:
grp= sns.distplot(df_info6['BR_C'], label= 'cluster 6', bins= 30, kde= True)
grp= sns.distplot(df_info7['BR_C'], label= 'cluster 7', kde= True)
plt.legend()
plt.show()

20-40대 서울 거주 여성을 클러스터링함

In [48]:
pt_specified= pt_bytsne.loc[(df_demo['IS_FEMALE']==1)&((df_demo['AGE_PRD']==20)|(df_demo['AGE_PRD']==30)|(df_demo['AGE_PRD']==40))&(df_demo['PRVIN']=='서울특별시'), :].reset_index(drop= True)
df_specified= df_item_consumed.loc[(df_demo['IS_FEMALE']==1)&((df_demo['AGE_PRD']==20)|(df_demo['AGE_PRD']==30)|(df_demo['AGE_PRD']==40))&(df_demo['PRVIN']=='서울특별시'), :].drop('ID', axis= 1).reset_index(drop= True)
df_specified.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 2554 entries, 0 to 2553
Columns: 149 entries, H&B선물세트 to 훼이셜케어
dtypes: float64(149)
memory usage: 2.9 MB
In [49]:
plot_results(pt_specified[[0,1]])
In [50]:
label_dict= {}
for i in range(10, 41, 2):
    gmm = GaussianMixture(n_components=i, covariance_type='full', random_state= 42).fit(pt_specified[[0,1]])
    label_dict[(2,i)]= pd.Series(gmm.predict(pt_specified[[0,1]]))
    plot_results(pt_specified[[0,1]], label_dict[(2,i)])
In [51]:
gmm = GaussianMixture(n_components=10, covariance_type='full', random_state= 42).fit(pt_specified[[0,1]])
label_dict[(2,i)]= pd.Series(gmm.predict(pt_specified[[0,1]]))
plot_results(pt_specified[[0,1]], label_dict[(2,10)])
In [52]:
plot_clusters(pt_specified[[0,1]], label_dict[(2,10)])
In [53]:
find_prefered_item(df_specified, label_dict[(2,10)])

for cluster 0:

item 가공식품 has different mean with total dataset; (3.8085240588030733, 0.00017855907703891364)
item 골프용품 has different mean with total dataset; (7.747799521935178, 3.3028682686205725e-13)
item 남성정장 has different mean with total dataset; (3.7142440627692577, 0.00026118070718228573)
item 시티웨어 has different mean with total dataset; (3.396846555905115, 0.0008163765024180523)
item 피혁잡화 has different mean with total dataset; (7.46102150568361, 2.161620614550045e-12)

for cluster 1:


for cluster 2:

item 가공식품 has different mean with total dataset; (7.768563116556986, 8.632857365934452e-14)
item 가공우유 has different mean with total dataset; (10.932506203841688, 1.4418737628213408e-25)
item 고급 has different mean with total dataset; (3.6250139587881316, 0.0003328805841116394)
item 과일야채 has different mean with total dataset; (2.315532928404998, 0.021146692635860914)
item 너트류 has different mean with total dataset; (2.752705443325415, 0.0061647106949745635)
item 봉투보증금 has different mean with total dataset; (2.5177112819550085, 0.012206704267510896)
item 스타킹 has different mean with total dataset; (2.7977924479536473, 0.005413844186700568)
item 육포 has different mean with total dataset; (2.020673025076723, 0.04403950809086476)
item 피혁잡화 has different mean with total dataset; (2.687392396619787, 0.007511856447438152)
item 화과자 has different mean with total dataset; (2.9009714137465146, 0.003937094623309042)

for cluster 3:


for cluster 4:

item 수족관용품 has different mean with total dataset; (2.1724952777779087, 0.030551809870899796)
item 유아용품 has different mean with total dataset; (3.144675680992624, 0.0018269084621517506)
item 차 has different mean with total dataset; (1.9726210078336626, 0.04938904825931994)
item 편의시설 has different mean with total dataset; (3.628890077832034, 0.00033344106994784837)
item 푸드코트 has different mean with total dataset; (2.5569910540373555, 0.01093327028319668)

for cluster 5:

item 생선초밥 has different mean with total dataset; (1.9974377964120547, 0.04649707887605989)
item 수입게 has different mean with total dataset; (2.1098169933841864, 0.03557269059083404)
item 수입돼지고기 has different mean with total dataset; (2.313819346608773, 0.021198363126358748)
item 스낵 has different mean with total dataset; (6.083232856171772, 2.5305524223346694e-09)
item 시즌잡화 has different mean with total dataset; (1.9681100431969147, 0.049835182248180995)
item 시즌행사 has different mean with total dataset; (2.291910591785781, 0.02246282242982383)
item 시티캐주얼 has different mean with total dataset; (2.1293223078884145, 0.033876210994879336)
item 애견용품 has different mean with total dataset; (3.7085599114743513, 0.0002432140340556605)
item 언더웨어행사 has different mean with total dataset; (2.346292086959509, 0.019473856842612283)
item 여성양말 has different mean with total dataset; (2.1643476221934765, 0.03102257892084076)
item 유가공 has different mean with total dataset; (4.176799503439039, 3.676654462083629e-05)
item 육포 has different mean with total dataset; (3.2019461298643233, 0.00148298106021081)
item 이온음료 has different mean with total dataset; (2.8716837953447407, 0.0043214033810854684)
item 이지캐주얼브랜드 has different mean with total dataset; (5.40271711345169, 1.0372225694841155e-07)
item 인스턴트선물세트 has different mean with total dataset; (1.9836201016303627, 0.048054319362406846)
item 일반욕실용품 has different mean with total dataset; (3.0399077772414724, 0.0025316577674734503)
item 일반헤어케어 has different mean with total dataset; (3.759343615986001, 0.00019808824505924697)
item 자몽 has different mean with total dataset; (2.759702699108514, 0.006078253879965392)
item 전통주 has different mean with total dataset; (4.9276889794774315, 1.2988736712915933e-06)
item 제습제 has different mean with total dataset; (2.290881091028924, 0.022502373213416367)
item 조리반찬 has different mean with total dataset; (3.615179759407444, 0.00034454655959113677)
item 즉석도정미 has different mean with total dataset; (2.0872165378389225, 0.03751798732885787)
item 차 has different mean with total dataset; (2.9186966835852233, 0.0037234957639257034)
item 탈취제 has different mean with total dataset; (2.4946654586407866, 0.013049228562835675)
item 튀김 has different mean with total dataset; (4.545475033002083, 7.561063943675574e-06)
item 파인애플 has different mean with total dataset; (2.7838426018297553, 0.0056458809156291045)
item 풋/핸드케어 has different mean with total dataset; (1.9786179514492492, 0.048622420426154146)
item 피자 has different mean with total dataset; (2.320390521273679, 0.020867300382270016)
item 해초 has different mean with total dataset; (2.5402600381080704, 0.011510233322031474)
item 호주산소고기 has different mean with total dataset; (6.935512641064557, 1.990469936554774e-11)
item 황태 has different mean with total dataset; (3.4600801560665273, 0.0006029738288972806)
item 훼이셜케어 has different mean with total dataset; (3.141371962646869, 0.001818618608937248)

for cluster 6:

item 가공식품 has different mean with total dataset; (5.3705029867517, 1.63090216167977e-07)
item 가공우유 has different mean with total dataset; (9.766798119220702, 4.404866944766555e-19)
item 골프용품 has different mean with total dataset; (3.3821321767296104, 0.0008449112387239279)
item 과자 has different mean with total dataset; (8.059892696900096, 4.11547958337076e-14)
item 선케어 has different mean with total dataset; (2.0616285642951326, 0.04030473330156683)
item 스타킹 has different mean with total dataset; (2.3012379060560617, 0.022194721261132807)
item 시즌행사 has different mean with total dataset; (2.014070182098718, 0.045057710239330556)
item 시티웨어 has different mean with total dataset; (2.020750761258648, 0.04435542901216915)
item 이온음료 has different mean with total dataset; (2.131447692886, 0.034093037270732034)
item 차 has different mean with total dataset; (2.774530815084145, 0.005945135435441225)
item 화과자 has different mean with total dataset; (2.4412500746132944, 0.015315700504341246)

for cluster 7:

item 아이메이크업 has different mean with total dataset; (3.3007929202767854, 0.001107475533667504)
item 피혁잡화 has different mean with total dataset; (3.3075187621068816, 0.0010537223588615871)

for cluster 8:

item VIDEOGAME has different mean with total dataset; (2.2472907225010292, 0.025573732936395262)
item 가공식품 has different mean with total dataset; (2.670729409889826, 0.008089590432169523)
item 레스토랑 has different mean with total dataset; (3.2725794824523646, 0.0012319311220810294)
item 수족관용품 has different mean with total dataset; (2.384283222560736, 0.01794567613256471)
item 스낵 has different mean with total dataset; (9.263231801993701, 1.7105364662509276e-17)
item 스포츠슈즈 has different mean with total dataset; (2.0799811277789675, 0.038653291618778424)
item 시즌행사 has different mean with total dataset; (2.981685276623306, 0.003174383381978067)
item 식사류 has different mean with total dataset; (3.9109944018321006, 0.00012256202385570964)
item 아동브랜드 has different mean with total dataset; (5.939606618178964, 1.1082059804010766e-08)
item 아로마 has different mean with total dataset; (2.5689943387818555, 0.010842560046393387)
item 여성양말 has different mean with total dataset; (3.018845303622133, 0.002829370191589617)
item 여성캐주얼브랜드 has different mean with total dataset; (2.7837343623653314, 0.005834402544196633)
item 유가공 has different mean with total dataset; (2.5986710764430376, 0.009970566878839517)
item 유아용품 has different mean with total dataset; (2.902493359160209, 0.004044797241726368)
item 이온음료 has different mean with total dataset; (2.24272789422134, 0.025867258164847917)
item 이유식/유아간식 has different mean with total dataset; (2.4639273689136805, 0.014513951074996171)
item 이지캐주얼브랜드 has different mean with total dataset; (7.87363323905385, 1.5574939498136062e-13)
item 일반욕실용품 has different mean with total dataset; (3.208927669134992, 0.0015145768532357435)
item 즉석구이김 has different mean with total dataset; (2.125818710359694, 0.03457272684281689)
item 즉석구이안주 has different mean with total dataset; (2.5665882122837544, 0.010895178647512915)
item 차 has different mean with total dataset; (2.421218100372205, 0.01620542553394315)
item 탈취제 has different mean with total dataset; (3.7764429455053463, 0.0002028697951768028)
item 파인애플 has different mean with total dataset; (2.001822241555434, 0.046457505993324925)
item 편의시설 has different mean with total dataset; (3.079677007353112, 0.0023204349647886918)
item 포장젓갈 has different mean with total dataset; (2.142034995326228, 0.03328777727172136)
item 푸드코트 has different mean with total dataset; (3.46063578569554, 0.0006473452442442861)
item 황태 has different mean with total dataset; (2.504657937816484, 0.012952353776990758)
item 훼이셜케어 has different mean with total dataset; (2.0537377164656214, 0.04099851034279835)

for cluster 9: